Be careful when opening your emails: a new type of computer virus uses artificial intelligence to write messages in a human style and spread to all your contacts.
A new computer virus, nicknamed “synthetic cancer”, has just emerged and it is proving to be particularly worrying. And for good reason: it uses artificial intelligence, in this case ChatGPT, to constantly metamorphose itself and spread insidiously, through electronic messaging. Above all, it is undetectable!
What makes this virus so formidable is its ability to constantly morph. ChatGPT automatically rewrites its code by changing variable names and slightly modifying its logic, which allows it to evade the detection mechanisms of traditional antivirus software. Once it has infected a computer, it uses the messaging system to send emails to all of its victim’s contacts. And since the messages are written by ChatGPT, in a human style, with the name of an existing contact, they appear completely legitimate.
The attack scenario is simple and clever. Imagine you receive an email from a friend inviting you to his birthday party with an attachment titled 80s_Nostalgia_Playlist. The message is so well-written that you have no reason to suspect that it is actually generated by a virus. By opening the attachment, you launch an executable program that allows the virus to install itself on your computer. From there, it starts infecting your own contacts in the same way, by digging through your messages and address book.
Fortunately, this scary virus is not roaming the wild: it is still in a lab. It was developed by David Zollikofer, a computer science student at ETH Zurich, and Ben Zimmerman, a cybersecurity researcher at Ohio State University, who shows how artificial intelligence can be exploited for malicious purposes. In doing so, the researchers want to highlight the urgent need to develop more effective defenses. Other experts, such as Alan Woodward from the University of Surrey, warn that we have only seen the tip of the iceberg when it comes to cyberattacks using artificial intelligence.
However, it is not all doom and gloom. Zollikofer believes that the same technologies can also be used to strengthen defenses against cyberattacks. By integrating artificial intelligence into security systems, it is possible to develop more sophisticated methods to detect and neutralize threats. The challenge is to stay one step ahead of the cybercriminals who exploit these new technologies.